Compression/Dilation Technique for Synchronizing Time-Series Signals Used to Detect Unwanted Electronic Components in Critical Assets Based on EMI Fingerprints

1. A method for detecting unwanted electronic components in a target asset, the method comprising: obtaining target electromagnetic interference (EMI) signals by monitoring EMI signals generated by the target asset while the target asset is running a periodic workload; generating a target EMI fingerprint from the target EMI signals; applying a compression/dilation technique to time-series signals in the target EMI fingerprint to achieve alignment with corresponding time-series signals in a reference EMI fingerprint to produce a synchronized target EMI fingerprint; and comparing the synchronized target EMI fingerprint against the reference EMI fingerprint to determine whether the target asset contains any unwanted electronic components; wherein the compression/dilation technique operates by compressing or dilating and then analytically resampling successive segments of the time-series signals in the target EMI fingerprint based on a moving time window to achieve a substantially optimal alignment with the corresponding time-series signals in the reference EMI fingerprint.

2. The method of claim 1 , wherein generating the target EMI fingerprint from the target EMI signals involves: performing a reference Fast Fourier Transform (FFT) operation on the target EMI signals to transform the target EMI signals from a time-domain representation to a frequency-domain representation; partitioning an output of the reference FFT operation into a set of frequency bins; constructing a target amplitude time-series signal for each of the frequency bins in the set of frequency bins; selecting a subset of frequency bins that are associated with the highest average correlation coefficients; and generating the target EMI fingerprint by combining reference amplitude time-series signals for each of the selected subset of frequency bins.

3. The method of claim 2 , wherein selecting the subset of frequency bins involves: computing cross-correlations between pairs of amplitude time-series signals associated with pairs of the set of frequency bins; computing an average correlation coefficient for each of the frequency bins based on the cross-correlations; and selecting a subset of frequency bins that are associated with the highest average correlation coefficients.

4. The method of claim 1 , wherein prior to obtaining the target EMI signals, the method further comprises generating the reference EMI fingerprint by: obtaining reference EMI signals, which are generated by a reference asset of the same type as the target asset while the reference asset is running the periodic workload, wherein the reference asset is certified not to contain unwanted electronic components; and generating the reference EMI fingerprint from the reference EMI signals.

5. The method of claim 4 , wherein comparing the synchronized target EMI fingerprint against the reference EMI fingerprint involves: computing a cumulative mean absolute error (CMAE) between time-series signals in the synchronized target EMI fingerprint and time-series signals in the reference EMI fingerprint; and comparing the CMAE against a threshold value to determine whether the target asset contains any unwanted electronic components.

6. The method of claim 4 , wherein comparing the synchronized target EMI fingerprint against the reference EMI fingerprint involves: feeding synchronized target amplitude time-series signals into an inferential model to produce estimated values for the synchronized target amplitude time-series signals, wherein the inferential model was previously trained based on time-series signals in the reference EMI fingerprint; performing pairwise differencing operations between actual values and the estimated values for the synchronized target amplitude time-series signals to produce residuals; and analyzing the residuals to determine whether the target asset contains any unwanted electronic components.

7. The method of claim 6 , wherein analyzing the residuals involves: computing a cumulative mean absolute error (CMAE) based on the residuals; and comparing the CMAE against a threshold value to determine whether the target asset contains any unwanted electronic components.

8. The method of claim 6 , wherein analyzing the residuals involves: performing a sequential probability ratio test (SPRT) on the residuals to produce SPRT alarms; and determining from the SPRT alarms whether the target asset contains any unwanted electronic components.

9. The method of claim 1 , wherein the periodic workload comprises one of: a square-wave-shaped workload; and a sinusoidal workload.

10. The method of claim 1 , wherein the target asset comprises one of: a computer system; and a utility system component.

11. A non-transitory, computer-readable storage medium storing instructions that when executed by a computer cause the computer to perform a method for detecting unwanted electronic components in a target asset, the method comprising: obtaining target electromagnetic interference (EMI) signals by monitoring EMI signals generated by the target asset while the target asset is running a periodic workload; generating a target EMI fingerprint from the target EMI signals; applying a compression/dilation technique to time-series signals in the target EMI fingerprint to achieve alignment with corresponding time-series signals in a reference EMI fingerprint to produce a synchronized target EMI fingerprint; and comparing the synchronized target EMI fingerprint against the reference EMI fingerprint to determine whether the target asset contains any unwanted electronic components; wherein the compression/dilation technique operates by compressing or dilating and then analytically resampling successive segments of the time-series signals in the target EMI fingerprint based on a moving time window to achieve a substantially optimal alignment with the corresponding time-series signals in the reference EMI fingerprint.

12. The non-transitory, computer-readable storage medium of claim 11 , wherein generating the target EMI fingerprint from the target EMI signals involves: performing a reference Fast Fourier Transform (FFT) operation on the target EMI signals to transform the target EMI signals from a time-domain representation to a frequency-domain representation; partitioning an output of the reference FFT operation into a set of frequency bins; constructing a target amplitude time-series signal for each of the frequency bins in the set of frequency bins; selecting a subset of frequency bins that are associated with the highest average correlation coefficients; and generating the target EMI fingerprint by combining reference amplitude time-series signals for each of the selected subset of frequency bins.

13. The non-transitory, computer-readable storage medium of claim 11 , wherein prior to obtaining the target EMI signals, the method further comprises generating the reference EMI fingerprint by: obtaining reference EMI signals, which are generated by a reference asset of the same type as the target asset while the reference asset is running the periodic workload, wherein the reference asset is certified not to contain unwanted electronic components; and generating the reference EMI fingerprint from the reference EMI signals.

14. The non-transitory, computer-readable storage medium of claim 13 , wherein comparing the synchronized target EMI fingerprint against the reference EMI fingerprint involves: computing a cumulative mean absolute error (CMAE) between time-series signals in the target EMI fingerprint and time-series signals in the reference EMI fingerprint; and comparing the CMAE against a threshold value to determine whether the target asset contains any unwanted electronic components.

15. The non-transitory, computer-readable storage medium of claim 13 , wherein comparing the synchronized target EMI fingerprint against the reference EMI fingerprint involves: feeding synchronized target amplitude time-series signals into an inferential model to produce estimated values for the synchronized target amplitude time-series signals, wherein the inferential model was previously trained based on time-series signals in the reference EMI fingerprint; performing pairwise differencing operations between actual values and the estimated values for the synchronized target amplitude time-series signals to produce residuals; and analyzing the residuals to determine whether the target asset contains any unwanted electronic components.

16. A system that detects unwanted electronic components in a target asset, comprising: at least one processor and at least one associated memory; and a detection mechanism that executes on the at least one processor, wherein the detection mechanism: obtains target electromagnetic interference (EMI) signals by monitoring EMI signals generated by the target asset while the target asset is running a periodic workload; generates a target EMI fingerprint from the target EMI signals; applies a compression/dilation technique to time-series signals in the target EMI fingerprint to achieve alignment with corresponding time-series signals in a reference EMI fingerprint to produce a synchronized target EMI fingerprint; and compares the synchronized target EMI fingerprint against the reference EMI fingerprint to determine whether the target asset contains any unwanted electronic components; wherein the compression/dilation technique operates by compressing or dilating and then analytically resampling successive segments of the time-series signals in the target EMI fingerprint based on a moving time window to achieve a substantially optimal alignment with the corresponding time-series signals in the reference EMI fingerprint.

17. The system of claim 16 , wherein while generating the target EMI fingerprint from the target EMI signals, the detection mechanism: performs a reference Fast Fourier Transform (FFT) operation on the target EMI signals to transform the target EMI signals from a time-domain representation to a frequency-domain representation; partitions an output of the reference FFT operation into a set of frequency bins; constructs a target amplitude time-series signal for each of the frequency bins in the set of frequency bins; selects a subset of frequency bins that are associated with the highest average correlation coefficients; and generates the target EMI fingerprint by combining reference amplitude time-series signals for each of the selected subset of frequency bins.

18. The system of claim 17 , wherein selecting the subset of frequency bins involves: computing cross-correlations between pairs of amplitude time-series signals associated with pairs of the set of frequency bins; computing an average correlation coefficient for each of the frequency bins based on the cross-correlations; and selecting a subset of frequency bins that are associated with the highest average correlation coefficients.

19. The system of claim 16 , wherein prior to obtaining the target EMI signals, the detection mechanism generates the reference EMI fingerprint by: obtaining reference EMI signals, which are generated by a reference asset of the same type as the target asset while the reference asset is running the periodic workload, wherein the reference asset is certified not to contain unwanted electronic components; and generating the reference EMI fingerprint from the reference EMI signals.

20. The system of claim 19 , wherein comparing the synchronized target EMI fingerprint against the reference EMI fingerprint involves: computing a cumulative mean absolute error (CMAE) between time-series signals in the synchronized target EMI fingerprint and time-series signals in the reference EMI fingerprint; and comparing the CMAE against a threshold value to determine whether the target asset contains any unwanted electronic components.